PROJECT SUMMARY/ABSTRACT Melanoma incidence rates continue to increase, whereas rates for most other cancers have declined. Emerging systemic therapies for patients who have progressed to metastatic melanoma are extending survival, but can be toxic, immensely costly and have variable efficacy. There are efforts to extend the use of new agents to earlier disease stages as adjuvant therapies. Better prediction of lethal melanoma at time of diagnosis is important to determine those who would benefit from adjuvant therapies and, conversely, avoid toxicity in those whose melanomas are unlikely to recur. Our central hypothesis is that it is possible to determine the combinations of genetic alterations and immune responses in tumor samples that define a melanoma likely to be ultimately lethal if not treated early. Conversely, we hypothesize that it will be possible to identify somatic profiles of the tumors that indicate a very low risk of recurrence and ultimate death of the patient, obviating the need for expensive, toxic treatments in such patients. We also hypothesize that host characteristics, including germline variants in multiple genes, will be associated with lethal melanoma. In Aim 1, we will test whether common somatic mutations in 104 genes/gene regions in tumors are associated with death from melanoma. These genes include known oncogenic drivers (BRAF, NRAS and TERT) and tumor suppressors (NF1, CDKN2A/CDKN2B, PTEN and TP53). We will also test the hypothesis that measuring the immune cell milieu will add information to survival prediction beyond conventional grading of tumor-infiltrating lymphocytes. We will then explore how best to combine the genetic alterations, immune measures and 2018 American Joint Committee on Cancer (AJCC) tumor stage to identify those melanomas highly likely to be ultimately lethal. In Aim 2, we will determine the extent to which the host characteristics of germline variants, phenotypic characteristics, sun exposure, age and gender differ among melanoma cases characterized as potentially lethal vs. nonlethal, leading to the identification of a subset of the population at high risk for lethal melanoma. This research will be undertaken using a resource that is uniquely suited to addressing these issues, the international, population-based Genes, Environment and Melanoma (GEM) Study, which collected epidemiologic data, pathology and tumor sections for a large cohort of incident primary cutaneous melanoma cases diagnosed in the year 2000. We have available a minimum of 10 years of follow-up of vital status and cause of death for all cases. GEM is an ideal resource because the end of the follow-up period precedes the targeted and immune therapy era, allowing us to study factors affecting the natural history of the disease. The vast majority of GEM cases (more than 85%) presented with local disease, and we project that at least 80% of the deaths will occur in cases initially presenting with local or regional disease. The results should enhance our ability to identify lethal melanomas at an early stage when treatments may be more effective and, conversely, to identify melanomas with a very low risk of recurrence for which unnecessary and potentially toxic treatments can be confidently avoided.